Statistical Modeling Software SEMS
The Simultaneous Equations Modeling System (SEMS) software provides the user with the capability of modeling in accordance with 2SLS and 3SLS methods. Statistical modeling of economic or marketing behavior has long necessitated the use of econometric methods designed to overcome some difficulties inherent in the use of classical regression analysis. The theoretical assumptions, under which the classical least-squares method provides unbiased estimators, are rarely attainable when modeling the relationship between economic or marketing variables. In many instances, the regressors (independent variables) are stochastically generated and cannot therefore be assumed to have ‘fixed’ values in repeated samples. Furthermore, the relationship under investigation may be widely divergent encompassing several simultaneous relations, which are expressed in a system of stochastic equations. In such a system, the regressants (dependent variables) are to be interactively determined among themselves and jointly determined along with the regressors. Under these circumstances, the attempt to apply an ordinary linear regression analysis in order to estimate the coefficients of the linear functions produces biased estimators.